Saagar Jha

https://federated.saagarjha.com/users/saagar

Website: https://saagarjha.com
Email: saagar@saagarjha.com
Hotel: Trivago

Public Key

npub1qacs7a8qxgwqsv595laqt8qh3xlmp2e6sq953mfkdflwu2up5j7qfy9wrv

NIP-05 Address

saagar_at_federated.saagarjha.com@momostr.pink

Profile Code

nprofile1qqsqwug0wnsry8qgx2z607s9nstcn0as4vagqz6ga5mx5lhw9wq6f0qpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds3d4tn6

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1qacs7a8qxgwqsv595laqt8qh3xlmp2e6sq953mfkdflwu2up5j7qfy9wrv

Show more details

Last Notes

2026-03-22 10:16:02 UTC

My security hot take for this week is that Google’s changes for sideloading on Android seem to strike a good balance between security and usability. This gives me hope the team is putting thought into maintaining the original dream of the platform rather than making a worse iOS

2026-03-06 13:31:01 UTC

Maybe I have only worked on discount bin browsers but there is no way any codebase that size has 10% of its crashes coming from hardware defects

2026-02-04 03:22:32 UTC

If you’re an developer for Apple’s platforms and were wondering where you rank in their list of priorities consider that they were apparently capable of writing docs and adding meaningful Xcode integrations all this time but they decided to do it to help AI models instead of you

2025-12-30 13:49:11 UTC

- reply

Except sudo, which has merged all 23 of his PRs without comment. I guess this means their code is either so bad that the LLM is finding all bunch of actual bugs, or that the maintainers aren’t reading his PRs closely. I’m not sure either makes me feel good about the project.

2025-12-30 13:46:52 UTC

- reply

There are no obvious tells and some of them do actually fix real (if often superficial problems) so he is being taken seriously by most of them, it’s just that half of them get merged and half are closed as per above because they’re wrong.

2025-12-30 13:44:32 UTC

So there’s this guy on GitHub that’s sending hundreds of PRs to completely unrelated open source projects to fix typos and stuff and I’m sure he’s using AI or something because half of them are unnecessary or wrong and when someone reviews the PR to point it out he just closes it

2025-09-16 18:31:59 UTC

Hi it’s me I’m popular demand
https://federated.saagarjha.com/media/74e58063-93b2-4aa5-ba83-13db3a54903b/file.jpeg

2025-09-09 17:47:16 UTC

Hmm did they add tensor cores to the GPU

2025-06-17 04:29:32 UTC

- reply

Depending on your relationship with the reporter and the project they’re filing bugs against, you might be able to provide additional insight or help move the report along if it gets stuck. Hopefully these tips help you snag some high quality feedback and turn them into fixes!

2025-06-17 04:25:12 UTC

- reply

There’s actually a process that tries to do this automatically, but it usually doesn’t work, and it takes a very long time to see results. This way, the reporter feels like you’re an advocate for their bug, even though you don’t need to do much at all. It’s a win-win!

2025-06-17 04:23:17 UTC

- reply

Once you’ve done this, you should put the bug on a list that you check every week so. You don’t need to do anything else, just check the status of the feedback and relay any requests for information back to the reporter directly so they can follow up.

2025-06-17 04:19:51 UTC

- reply

Sometimes you don’t have enough information to move a bug. That’s easy to resolve too! Just ask the reporter for more details. If you tell them why you need it, they’ll often help out. This can save them and the screeners a ton of time and get feedback in earlier.

2025-06-17 04:16:59 UTC

- reply

It’s a common misconception that feedback goes to the right category automatically! Unfortunately, third party reporters do not have access to all the components you do. The more a bug needs to move around, the more likely it’ll get lost or misunderstood. You can fix that!

2025-06-17 04:14:37 UTC

- reply

Thankfully, this is easy to fix. All you need to do is paste the feedback number into Radar, then read the report. Once you understand what it’s asking for, just move it along into the right component. Then: reply to the reporter with “the right people are looking at it”.

2025-06-17 04:12:14 UTC

- reply

First off, you’ve got to be engaged in the bug reporting process. Did you sit on a feedback number someone sent you? That’s a big red flag. Nobody is going to commit to putting in effort if you can’t do the bare minimum yourself. Filing bugs is a two-way process!

2025-06-17 04:09:55 UTC

It’s beta season, which means there’s a lot of advice for third party developers posted online on how to file good feedback. But what if you’re an Apple engineer? Is there anything you should be doing besides asking for people to send in reports? Here’s a handy checklist!

2025-06-09 23:32:18 UTC

Make a backup, they said
https://federated.saagarjha.com/media/b9942183-aceb-4d5a-ba50-0be2ed164602/file.png

2025-05-26 07:01:49 UTC

Person A: I want to shoot myself in the foot
Me: Can you explain what you actually want to do
Person B: Here’s how to shoot yourself in the foot
Person A: Oh, that works. Thanks, Person B
Being in any technical space means this conversation happens a lot more often than I’d like

2025-05-01 10:33:01 UTC

- reply

> Mr. Onak commented that “if we want to ‘scare’ users a bit, i like the addition of ‘out’ because it raises questions and hesitancy haha. out? out where? omg what do i do?”
Well then Mr. Onak, what *will* you do?

2025-05-01 10:31:08 UTC

- reply

> One employee further wrote, “to make your version even worse you could add the developer name rather than the app”. To that, another responded “ooh - keep going.”
"At Apple, design is core to everything we do. Our team works hard to make things worse.”

2025-05-01 10:25:54 UTC

- reply

In a hilarious twist Gonzalez literally takes a screenshot of Apple’s own Human Interface guidelines to call them out
https://federated.saagarjha.com/media/10168500-c57d-4ba1-bb1c-bc1add9a607b/file.png

2025-05-01 10:19:12 UTC

- reply

> Confusion is not a security risk. Given the lack of any document identifying this alleged concern, the Court finds these justifications pretextual
Gonzalez is not having it lmao

2025-05-01 10:15:52 UTC

- reply

> In its notice of compliance and at the May 2024 hearing, Apple claimed that restrictions on
link placement protect against “security risks.” […] Again, Apple attempted to mislead.
Ooh it’s getting spicy

2025-05-01 10:12:35 UTC

- reply

> Apple hired the Analysis Group (“AG”), …, which (1) “[e]stimate[d] the value of services provided by the Apple ecosystem to developers,”
> Nonetheless, AG “estimated”
> AG “concluded”
The quotes, of course, indicate how much the court trusts this third party analysis group

2025-05-01 10:04:05 UTC

- reply

Nobody is really talking about the Epic side today but the court does confirm that they breached the DPLA and need to pay attorney’s fees (which I don’t think they really care about at this point)

2025-05-01 10:02:50 UTC

- reply

Loving the use of “supracompetitive” going to use it all the time now

2025-05-01 10:01:01 UTC

- reply

> Thus the Court found:
> In retail brick-and-mortar stores, consumers do not lack knowledge of options. Technology platforms differ.
brb going to annihilate all the Hacker News posters talking about how this is actually the same as Walmart

2025-05-01 09:57:34 UTC

Get in losers, we’re reading Judge Gonzalez’s ruling
(Note: you can get the zingers elsewhere, I am going to assume you read those already. If you haven’t, you should, they are *really* funny.)

2025-01-04 09:58:07 UTC

- reply

I mean I am sure your business conduct training has different things to say about paying bribes

2025-01-03 18:42:21 UTC

- reply

I mean like all the other frameworks

2025-01-03 13:38:29 UTC

- reply

I am sure random bits of the OS would probably get real upset if things suddenly become big endian as well

2024-12-02 02:23:57 UTC

- reply

I think the bigger issue here is to solve pointer chasing first because GPUs are even worse than CPUs at it

2024-11-14 22:13:17 UTC

- reply

I like how it’s easy to figure out exactly where this is even with the censorship lol

2024-11-14 08:21:05 UTC

- reply

The interesting part is that Rosetta apparently doesn’t require TSO anymore

2024-11-12 00:22:14 UTC

- reply

Paying people in stock typically does not align their actions with what benefits the company

2024-11-07 03:17:45 UTC

Someone (I can’t remember who) asked at one point whether Apple actually documented their TSO extension. Until recently the best source was a rare post to LKML by an Apple engineer, but they’ve actually published official Linux patches to turn it on now: https://developer.apple.com/documentation/virtualization/accelerating_the_performance_of_rosetta

2024-10-30 08:57:03 UTC

- reply

I like how she demonstrates the power of M4 by showing it running Slack

2024-10-30 08:55:21 UTC

- reply

Make this you cowards
https://federated.saagarjha.com/media/5184e04c-a6ff-49b4-b548-ad7612afbadb/file.png

2024-10-30 08:52:04 UTC

- reply

Do they not do food at Caffè Macs on Bandley anymore or is it like just a prop now

2024-10-30 08:48:01 UTC

- reply

Making Apple Park smol for mini Mac is kind of cute

2024-10-30 08:46:45 UTC

I am told that there is an introduction video for Mac mini which is basically like a slice of a keynote so I watched it and will make you suffer some livetweets*
*You can tell these are not livetweets because they have images–I can’t keep up with alt text during a livestream

2024-10-28 02:43:51 UTC

- reply

11:49?

2024-10-20 11:51:58 UTC

- reply

It’s kind of amusing there is a permission for network access but no toggle for it

2024-10-06 01:21:47 UTC

- reply

They’re missing you anyway. Go pay them a visit

2024-09-26 01:00:43 UTC

- reply

You beat my prediction for how long it would be before you replied by a solid 20 minutes ;)

2024-09-26 00:53:42 UTC

One of the biggest concerns I have with relying on indie iOS software is that I have to hedge against the author being hired by Apple and never being able to update the app ever again

2024-09-20 20:24:19 UTC

- reply

Maybe you should do a better job of tracking what is actually being used

2024-08-09 02:55:11 UTC

- reply

I have this sentiment about a lot of projects and it’s very difficult to express it without being yelled at as not being thankful for free labor

2024-07-02 01:49:42 UTC

- reply

Apple doesn’t generally obfuscate their code but yes this is what I use it for

2024-07-01 22:38:56 UTC

I’ve been busy the last couple weeks and dropped the ball on talking about this when it came out. The quality of the new decompiler is really good–easily letting it compete in the big leagues now (while other, unnamed, decompilers have mostly remained stagnant or regressed).
https://infosec.exchange/@binaryninja/112645012576820209

2024-06-08 21:40:25 UTC

Don’t want to jinx it but with WWDC right around the corner I have to say this is the first year that a network Time Machine Backup has actually lasted all the way to June without corrupting itself and forcing me to wipe it at least once